Dissertation/Thesis Abstract

An algorithmic process for graphical simulation of manipulator motion trajectories
by Bhangaonkar, Praneel, M.S., Northern Illinois University, 2012, 171; 1513409
Abstract (Summary)

Manipulator motion trajectory controls generally start with defining the task structures. Task structures, sequences of manipulator motions required to complete a task, are transformed into task motion trajectories. These motion trajectories are implemented using Joint Motion or Cartesian Motion Scheme. In the Joint Motion Scheme the End of Manipulator (EOM) positions are translated into joint variables at the servo speed. This translation of EOM position into joint variables may be done on-line or off-line. The advantage of this motion scheme is that obstacles and singularities along a motion trajectory do not pose a problem as it is limited only by joint accelerations and maximum joint velocities. The main drawbacks of this motion scheme are the requirement of a pre-defined EOM path and/or evaluation overhead time.

The Cartesian Motion Scheme on the other hand is mostly, but not limited to, used in off-line controls. The main advantage of this scheme is that it does not require evaluation of the absolute joint parameters as in Joint motion scheme. Instead it uses differential joint actuator signals for any differential EOM motion between differential evaluation points. The major issue with this approach is that the EOM singularity and/or unexpected obstacles may be encountered in this on-line scheme. The Cartesian Motion Schemes are used in on-line mode with a risk of singularity and path obstacles solely with the main purpose of a shorter implementation time.

The governing equations for EOM motions in Cartesian Motion Scheme are well established. However for specific applications, after generating motion parameters using this scheme, the effectiveness and accuracy of this scheme need to be simulated before implementation. It was found in the literature survey that there was hardly any work done in regards to testing these models. The main goal of this thesis is to develop an algorithmic process to simulate motions for a specific time increment and to compare them with the expected path in terms of accuracy and possible EOM singularity. To achieve this, a typical object pick and place application was used and the motion trajectory was modeled using the Cartesian motion scheme. A program was developed in MATLAB software to generate the Drive Transforms and the joint differential variables were calculated using corresponding differential transformations. A graphical simulation was obtained based on the results for each of the joint variable and also for each of the joint differential variable.

Indexing (document details)
Advisor: Ganesan, Sengoda
Commitee: Song, Shin-Min, Woo, Peng Yung
School: Northern Illinois University
Department: Mechanical Engineering
School Location: United States -- Illinois
Source: MAI 50/06M, Masters Abstracts International
Subjects: Mechanical engineering
Publication Number: 1513409
ISBN: 978-1-267-42796-0
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy